A method for the determination of ethyleneurea in ... - ACS Publications

samples in volumes of 1.0 mL or less before extraction. ... Alumina used for adsorption chromatography (Neutral ... fitted, and the tube heated at 75 ...
1 downloads 0 Views 402KB Size
J. Agric. Food Chern., Vol. 26, No. 6, 1978

Ethyleneurea Determination in Food

1325

A Method for the Determination of Ethyleneurea in Foods as the Pentafluorobenzamide William H. Newsome A method was developed for the determination of ethyleneurea in foods at levels of from 0.01 to 5 ppm. Residues were extracted with acetone and the extract alkylated to remove ethylenethiourea (ETU). After partial cleanup on alumina, ethyleneurea was pentafluorobenzoylated and the product purified on a column of silicic acid. Quantitation of the pentafluorobenzamide was carried out by high-pressure liquid chromatography with confirmation by GLC and mass spectrometry. Recoveries from five commodities were generally greater than 80%. The method was unaffected by zineb, whereas 3% of added E T U was converted to ethyleneurea. A preliminary survey of three commodities indicated that approximately 33% of the samples contained ethyleneurea. Ethyleneurea (2-imidazolidinone) is a reaction product of the ethylenebis(dithi0carbamates) (EBDCs) and represents a major metabolite of these fungicides. It has been identified in plants treated in the greenhouse with [14C]zineb (Vonk, 1976) or outdoors with [3H]mancozeb (Lyman, 1971) or [14C]maneb,[14C]zineb,[14C]nabam,and [14C]ETU (Nash, 1976). Ethylenethiourea (ETU), a transient metabolite of the EBDCs, is a precursor of ethyleneurea in vitro in the presence of light and photosensitizers (Ross and Crosby, 1973; Cruickshank and Jarrow, 1973) and in vivo in plants (Hoagland and Frear, 1976; Nash, 1976). When fed with nitrite, ethyleneurea has been found carcinogenic in rats (Sander and Burkle, 1971). Although nitrosoethyleneurea is relatively unstable in aqueous solution (Lee et al., 1977) and thus unlikely to be found in foods, nitrosation of ethyleneurea has been reported to occur readily (Sander, 1972). Nitrosoethyleneurea is mutagenic in bacteria (Lee et al., 1977) and carcinogenic in rats (Druckrey et al., 1967). Since ethyleneurea constitutes a potential hazard to health and residue methods have not been described in the literature, the present method was developed to permit its determination in foods. T h e analytical procedure involves extraction and preliminary cleanup on alumina, followed by derivatization to form the pentafluorobenzamide. T h e derivative combines good absorption a t 254 nm for the detection of residues by high-pressure liquid chromatography (LC), with sufficient volatility and electron-capture response to permit its subsequent confirmation by gas-liquid chromatography (GLC). In addition, the derivative may be confirmed by high-resolution mass spectrometry after isolation by LC. EXPERIMENTAL SECTION

Materials. Ethyleneurea was purchased from Eastman Organic Chemicals, Rochester, N.Y., and was crystallized from acetone to give white needles, mp 131-132 "C (lit. 131-132 "C, Duschinsky and Dolan, 1946). Solutions for fortification were prepared in methanol and added to samples in volumes of 1.0 mL or less before extraction. Pentafluorobenzoyl chloride was obtained from PCR Research Chemicals, Inc., Gainesville, Fla. Pentafluorobenzoylethyleneurea standard was prepared by refluxing ethyleneurea (215 mg, 2.5 mmol) with pentafluorobenzoyl chloride (1.15 g, 5 mmol) in benzene (10 mL) Food Research Division, Bureau of Chemical Safety, Food Directorate, Health Protection Branch, Health and Welfare Canada, Tunney's Pasture, Ottawa K1A OL2, Ontario, Canada. 002 1-8561/78/1426-1325$0 1.OO/O

for 2 h. The reaction mixture was concentrated to 2-3 mL and applied to a 1.5 X 9 cm column of silicic acid (Woelm, 100-200 pm) in benzene. After washing with benzene (5 mL), the column was eluted with 10% ethyl acetate in benzene and the 55-135-mL fraction collected. The solvent was removed and the derivative crystallized from benzene-hexane. The crystals melted at 165-167 "C and gave a mass spectrum as shown in Figure 1. Alumina used for adsorption chromatography (Neutral Alumina AG-7, 100-200 mesh) was purchased from BioRad Laboratories, Richmond, Calif., and was deactivated to activity I11 by the addition of 6% water. Silicic acid (Woelm, Activity I, 100-200 pm) was supplied by ICN Pharmaceuticals Inc., Cleveland, Ohio, and was used without deactivation. Analytical Procedure. Samples of blended crop (5.0 g) were homogenized with acetone (50 mL) in a Sorvall Omni Mixer. T h e homogenate was filtered through Whatman No. 1 paper in a Buchner funnel using slight negative pressure. The filtrate was transferred to a 250-mL round-bottomed flask and methyl iodide (0.1 mL) added. After 30 min a t room temperature the solvent was evaporated to slightly less than 1 mL on a rotary evaporator. The residue was transferred to a 5-mL volumetric flask with methanol (4.0 mL) and made to the mark with water. Ethyleneurea was dissolved in 20% aqueous methanol to give a solution containing 1 pg/mL and an aliquot (1.0 mL) was carried through the remainder of the procedure as reference standard. Alumina adsorption columns were prepared by adding neutral alumina, activity I11 (4.0 g) to 50% methanol in acetone (15 mL) in a chromatography tube. After settling the adsorbent, fines and supernatant were aspirated. An aliquot (1.0 mL) of extract or standard was added to the column which, after adsorption of the sample, was eluted with 50% methanol in acetone (10 mL). The eluate was collected in a 15-mL centrifuge tube and evaporated to dryness under a stream of nitrogen in a 50 "C water bath. A freshly prepared solution of pentafluorobenzoyl chloride (20 pL) in acetone (0.5 mL) was added, an air condenser fitted, and the tube heated at 75 "C for 1h. After removal of the solvent with a stream of nitrogen, the samples were taken up in dichloromethane (2 mL). Silicic acid adsorption columns were prepared by adding silicic acid (2.0 g) to dichloromethane (15 mL) in a glass column. The excess solvent was run through the column and the sample added. After adsorption of the sample, the column was washed with dichloromethane (20 mL) and elution begun with 15% acetonitrile in dichloromethane (20 mL). The last 10 mL of this eluate was collected and evaporated under a stream of nitrogen to dryness. The sample was then taken up in an appropriate volume of C 1978 American Chemical Society

Newsome

J. Agric. Food Chem., Vol. 26, No. 6, 1978

1326

F

i t-

MW 28C

w

601

w 1

40

$1 z

i

a 204 i

Figure 2. Reaction scheme for the derivatization of ethyleneurea with pentafluorobenzoyl chloride. I

I ll

W C

2

a

N L W

m

0 E 0

Figure 1. Mass spectrum of 1-pentafluorobenzoyl-2-imidazo-

lidinone; ionization voltage 70 e\'.

3 -1 i L

15% acetonitrile in dichloromethane and an aliquot (50 FL) injected onto the liquid chromatograph. H i g h - p r e s s u r e Liquid C h r o m a t o g r a p h y (LC). Samples were injected onto a 2.2 mm x 50 cm Aerograph Micro Pak Si-10 column by a Valco loop injector fitted with a 50-pL sample loop. The column was eluted with 3% isopropyl alcohol in dichloromethane delivered a t a flow rate of 0.5 mL min-' by an Aerograph Model 4000 constant pressure pump. The pentafluorobenzamide was detected and quantitated by absorption a t 254 nm on a Waters Model 440 absorbance detector coupled to a 1-mV Honeywell recorder. For confirmation, the fraction with a retention time of the pentafluorobenzamide was collected in a 6 X 50 mm tube and evaporated to dryness with a stream of nitrogen. The residue was taken up in benzene and an aliquot (5 pL) analyzed by gas-liquid chromatography. For mass spectral confirmation the residue was dissolved in dichloromethane (30 pL) and an aliquot ( 2 gL) placed on the probe. Gas-Liquid Chromatography (GLC). Determinations were carried out on a Hewlett Packard Model 5700 A fitted with a 63Nielectron-capture detector and a 6 ft x 4 mm i.d. glass column packed with 670 QF-1 and 4% SE-30 on 80-100 mesh Supelcoport. Argon-methane (95:5)carrier gas was supplied to the column at a flow rate of 35 mL min-'. The column was operated a t 205 "C, while the injection port and detector were maintained at 200 and 300 "C, respectively. Under these conditions, and with routine working attenuation, approximately 0.2 ng of pentafluorobenzamide produced 50% full-scale deflection. Mass Spectrometry. Eluate from the LC was analyzed on a Varian Mat 311.4 double-focusing spectrometer using single ion monitoring of the molecular ion ( m / e 280.03) with a resolution of 5000. The sample was introduced a t 25 "C and heated to 200 "C within 20 s. By this procedure, 0.1 ng of pentafluorobenzamide corresponding to a sample containing 0.01 ppm ethyleneurea produced an ion current 70% of full-scale deflection. Samples determined by LC, GLC, or mass spectrometry were quantitated by comparison of the peak heights to that produced by a standard of ethyleneurea carried through the procedure. RESULTS AND DISCUSSION Derivatization of ethyleneurea was considered a necessary step in its determination, since it is insufficiently volatile for GLC and is devoid of chromophores which would permit its detection after LC. Previous work (Greenhalgh and Weinberger, 1965) had shown ethyleneurea amenable to acylation with various reagents although yields were often low. This approach was inves-

a

i-

z

W

a

LL

0 C 2

W

>-

0

30

60

90

120

REACTION TIME ( M I N )

Figure 3. Time course of reaction of pentafluorobenzoyl chloride with ethyleneurea in acetone at 75 "C. Ethyleneurea (1pg) was heated for various time intervals with pentafluorobenzoyl chloride (20 pL) in acetone (0.5 mL). Yield of pentafluorobenzamide was determined by LC. Table I. Effect of Methyl Iodide o n the Conversion of Various Levels of ETU to Ethvleneurea"

ETU added,

ethyleneurea found, ppm

ppm

without CH,I

with CH,I

0.10 0.50

0.016 0.067

1.0

0.113

0.004 0.010 0.025

ETU was added to tomato homogenate and the homogenate analyzed for ethyleneurea. a

tigated with p-nitrobenzoyl chloride and pentafluorobenzoyl chloride, the latter being selected because of its favorable physical characteristics. The reaction with pentafluorobenzoyl chloride occurs as shown in Figure 2, giving the monoacylated product. A yield of approximately 60% was obtained after 1 h of reaction as shown in Figure 3. The yield could not be increased by longer reaction times, increases in reagent concentration, or change in solvent to dichloromethane, benzene, acetonitrile, or dimethylformamide. The addition of powdered anhydrous sodium carbonate or triethylamine was without affect. Because of the incomplete reaction, it was necessary to carry a standard of ethyleneurea through the procedure for quantitation purposes. A preliminary cleanup on alumina was found necessary to avoid decreases in yields of derivative during the reaction step. To obtain quantitative recovery of ethyleneurea from the alumina, it was important to control the amount of water in the sample and to deactivate the alumina by the addition of 6% water to give activity I11 (Brockmann). Ethylenethiourea (ETU), an EBDC metabolite which may occur concomitantly with ethyleneurea, was found to be partially converted to ethyleneurea by the analytical method. As shown by the data in Table I, alkylation reduced the conversion from an average of 13% to an

Ethyleneurea Determination in Food

J. Agric. Food Chem., Vol. 26, No. 6, 1978

Table 11. Effect of 5.7 ppm Zineb o r 0.1 ppm ETU on the Recovery of Ethyleneurea from Tomato

Table 111. Recoveries of Ethyleneurea Added t o Various Commodities

ethyleneurea found, PPm ETU zineb

ethyleneurea added, ppm 0 0.103 0.513

ethyleneurea adtoded, ppm mato

0 0.105 0.458

0 0.102 0.505

1327

0.01 0.05 0.10 0.51

1.0 5.1

93 79 93 82 84 99

ethyleneurea recovered,a % potato

carrot

apple

lettuce

91 87 86 85 103 90

91b 81b 80 80 83 85

46b

91b 72b 95b 87 92 92

62b

81 94 96 86

Recoveries are calculated relative t o a standard of ethyleneurea carried through the procedure simultaneously. Results are the means of duplicate determinations. Values determined by gas-liquid chromatography. Table IV. Ethyleneurea Content of Three Commercial Commodities 0

8

4

1

2

s

8

4

1

2

0

8

4

mean of positive positive samples, samples ppm

I2

1

B

commodity spinach

0 0 5 ppm EU

I

tomato

fresh processed fresh processed

potato a I

13/23 Oil0 3/10 8/15 4/15

0.124

0.020 0.018 tr

range of posi. tive samples, PPm 0.026-0.796 0.011-0.037 0.010-0.025 [0.004-0.001~

Tentative assignment.

on LC precluded an accurate measurement of peak heights. A preliminary survey of spinach, tomatoes, and potatoes purchased in the Ottawa area was carried out to obtain an indication of the presence of ethyleneurea in foods of commerce. As shown by the data in Table IV approximately 33% of the 73 samples contained ethyleneurea. Representative positive samples of each commodity were confirmed by mass spectrometry. The levels of ethyleneurea found in potatoes are considered tentative, since recovery data was not obtained below 0.01 ppm.

_I____

0

4

8

2 0 RETE\'

4

OI,

' 2 0 YE ' Y \UTES) 8

4

8

I2

F i g u r e 4. Chromatograms obtained from derivatized extracts of potato fortified with various amounts of ethyleneurea. Chromatograms in A resulted from LC with detection at 254 nm. Eluate from the silicic acid cleanup was evaporated to dryness, the residue taken up in 1.0 mL of solvent, and a 50-wL aliquot representing the equivalent of 100 mg of sample injected. Fractions from samples chromatographed in A were collected, evaporated, and dissolved in benzene (100 wL for blank and 0.01 ppm, and 200 pL for 0.05 pprn). A 5-pL aliquot was injected into a GLC with electron-capture detector to give the tracings shown in B.

average of 3% by weight. The data in Table I1 show that neither E T U nor zineb affected the recovery of ethyleneurea, nor was any degradation of zineb to ethyleneurea evident. In practice, levels of ETU found on raw foods are generally less than 0.1 ppm (Newsome, 1976; Pease and Holt, 1977; Ripley et al., 1978) and such amounts would result in ethyleneurea values less than 0.01 ppm. However, cooked foods may contain E T U a t levels considerably in excess of 0.1 ppm (Newsome, 1976; Onley et al., 1977; Report of the IUPAC Commission on Pesticide Terminal Residues, 1977). In the latter instance, significant amounts of ethyleneurea would be generated by the method, the magnitude of which may be determined by analysis for ETU. The recoveries of ethyleneurea added to various commodities are given in Table I11 and are generally greater than 80% from 0.01 to 5.1 ppm. Typical chromatograms are shown in Figure 4. Somewhat lower recoveries were obtained from apple of levels below 0.1 ppm. This was attributed to syrupy impurities present after chromatography on alumina which may have reduced the derivatization. Although rigorous recovery studies were not carried out at levels below 0.01 ppm, 0.005 ppm has been readily detected by GLC and the identity confirmed by mass spectrometry. With some commodities a t concentrations below 0.1 ppm, accurate quantitation was possible only by GLC, since the rapidly declining baseline obtained

ACKNOU'LEDGMENT

The author wishes to thank Luz Panopio for technical assistance and Walter Miles for mass spectral determinations. LITERATURE CITED Cruickshank, P. A., Jarrow, H. C., J . Agric. Food Chem. 21, 333 (1973). Druckrey, H., Preussmann, R., Ivankovic, S., Schmahl, D., Z. Krebsforsch. 69, 103 (1967). Duschinsky, R., D o h , L. A,, J . Am. Chem. Soc. 68,2350 (1946). Greenhalgh, R., Weinberger, M. A., Can. J . Chem. 43, 3340 (1965). J. Agric. Food Chem. 24, 129 (1976). Hoagland, R. E., Frear, D. S., Lee, K., Gold, B., Mirvish, S. S., Mutation Res. 48, 131 (1977). Lyman, W. R., Pure A p p l . Chem., Suppl., 243 (1971). Nash, R. G., J . Agric. Food Chem. 24, 596 (1976). Newsome, W. H., J . Agric. Food Chem. 24, 999 (1976). Onley, J . H., Giuffrida, L., Ives, N. F., Watts, R. R., Storherr, R. W., J . Assoc. Off. Anal. Chern. 60, 1105 (1977). Pease, H. L., Holt, R. F., J . Agric. Food Chem. 25, 561 (1977). Report of the IUPAC Commission on Pesticide Terminal Residues, Pure Appl. Chem. 49, 675 (1977). Ripley, B. D.. Cox, D. F., Wiebe, J . , Frank, R., J . Agric. Food Chem. 26, 134 (1978). Ross, R. D., Crosby, D. G., J . Agric. Food Chem. 21,335 (1973). Sander, J., Burkle, G., 2. Krebsforsch. 75, 301 (1971). Sander, J., in "N-Nitroso Compounds Analysis and Formation", IARC Scientific Publication No. 3, 1972 p 97. Vonk, J. W., Med. Fac. Landbouicx. Rijksuniu. Gent, 41, 1383 (1976). Received for review June 1, 1978. Accepted August 16, 1978.